addressalign-toparrow-leftarrow-leftarrow-right-10x10arrow-rightbackbellblockcalendarcameraccwcheckchevron-downchevron-leftchevron-rightchevron-small-downchevron-small-leftchevron-small-rightchevron-small-upchevron-upcircle-with-checkcircle-with-crosscircle-with-pluscontroller-playcredit-cardcrossdots-three-verticaleditemptyheartexporteye-with-lineeyefacebookfolderfullheartglobe--smallglobegmailgooglegroupshelp-with-circleimageimagesinstagramFill 1languagelaunch-new-window--smalllight-bulblightning-boltlinklocation-pinlockm-swarmSearchmailmediummessagesminusmobilemoremuplabelShape 3 + Rectangle 1ShapeoutlookpersonJoin Group on CardStartprice-ribbonprintShapeShapeShapeShapeImported LayersImported LayersImported Layersshieldstar-shapestartickettrashtriangle-downtriangle-uptwitteruserwarningyahooyoutube

Re: [LACloud-Computing] Elastic Filesystems

From: HQTran
Sent on: Monday, January 25, 2010, 1:27 PM
Darren - I have no experienced with this myself as I have only noticed this in newsletter announcement from Dreamhost - http://ceph.newdream.net/

This may or may not answer some of your needs I'd appreciate feedback on what you find should you investigate further.

-Hung

On Mon, Jan 25, 2010 at 1:20 PM, Darren <[address removed]> wrote:
I have been scouring the interwebs looking for an elastic filesystem that can be installed (non-S3) on my own gear. ?

I have probably read about or reviewed 50 different packages, but have yet to find any with the combination of features I am looking for. ?I'd love to know if anyone on the list knows of a project that fits the bill, or at least has similar requirements:

1) Master-less: ?The first node is a storage node. ?There are no 'master nodes' and therefore one less SPOF. ?Adding nodes increases total storage in a near-linear fashion. ?Most likely a Gossip protocol is used to achieve inter-node communication.
?
2) Redundant: Data is automatically replicated to achieve redundancy. ?Ideally these policies would be rack/colo aware

3) Elastic: You can add and remove nodes as needed. ?Nodes crashing are handled transparently. ?Data is rebalanced automatically in the background to ensure #2. ?libkatema or similar used here.

4) FUSE Support: A FUSE adapter is available - recognizing it may not be the most performant interface in all cases. ?A windows adapter would also be ideal - a WebDAV interface might be sufficient.

5) Open Source: Ideally written in a language more prevalent than modula

The closest thing I have found so far is Cassandra:

which is in use at Facebook, Digg, Twitter, etc. ?However, this is a document store, not a filesystem/filestore, which basically means it does more than what is needed, but also there are some filesystem semantics that are missing.

If I can't find what I'm looking for, I'm tempted to begin looking into a CassandraFS implementation that would leverage much of the core of Cassandra, but fill the gaps mentioned above.

Here is a short list of some of the projects I've looked at.
HDFS - Hadoop FS
Lustre
Gluster
MogileFS
Terrastore
Dokan - FUSE for windows
XtreemFS
Cassandra
Project Voldemort

I don't know what I don't know on this one - all leads appreciated.

Thanks,
Darren










--
Please Note: If you hit "REPLY", your message will be sent to everyone on this mailing list ([address removed])
This message was sent by Darren ([address removed]) from LA Cloud Computing.
To learn more about Darren, visit his/her member profile
To unsubscribe or to update your mailing list settings, click here

Meetup Inc. PO Box 4668 #37895 New York, New York[masked] | [address removed]

People in this
group are also in: